September 27, 2010
The Structural Unemployment Story  

Economics textbooks teach that there are three kinds of unemployment: frictional (when people voluntarily change jobs), cyclical (when downturns throw people out of work), and structural (when workers’ skills or locations don’t match available jobs). Obviously we’re now experiencing primarily cyclical unemployment. There was a housing bubble. It burst. That and a dramatically declining stock market wiped out a lot of wealth. So people now spend less. Unemployment accordingly has risen. Rocket science it isn’t.

Yet some economists and commentators would like to believe, and therefore do believe, that current unemployment is structural, caused primarily not by a deficiency of demand but by a mismatch between workers’ skills and the needs of the economy. David Brooks writes, “Today’s economic problems are structural, not cyclical. We are in the middle of yet another jobless recovery. Wages have been lagging for decades. Our labor market woes are deep and intractable.” (For sensible perspectives on the issue, see this, this, this, this, and this.)

As usual with Brooks, he provides no evidence, nor wastes time telling us what would need to be true if the structural unemployment story were true. So let me fill that in. What would need to be true is that depression in some industries is matched by exuberance (hiring, high wages, overtime) in others. After all, there’s no shortage of demand, just a change in the distribution of demand. So we wouldn’t see this picture:

 Yet the data here are real. The decline in employment is broad, not sectoral. The overall picture is this:

The above graph almost looks too patterned. And indeed some proponents of the structural unemployment story have seized on the sharp decline in construction employment to argue that the skills of construction workers no longer match those needed by the economy. But as Lawrence Mishel, Heidi Shierholz, and Kathryn Edwards note, in an important paper released last week by the Economic Policy Institute, they can’t have examined the labor market data too closely. For example, for the structural unemployment story to make sense, unemployed construction workers would need to comprise a disproportionately large share of the long-term unemployed. They don’t. Construction workers are 12.4% of current unemployment and 12.5% of long-term unemployment. Construction workers are thus as capable as most others of finding new work. There’s no mismatch between the skills of such workers and the needs of the economy.

Mischel, Shierholz, and Edwards also note that if unemployment were primarily structural, the ratio of job seekers to job openings would be about the same as in normal times—between 1 and 2. Currently it’s about 5:

The structural unemployment story also presumes, as the authors note, that “that millions of workers are now inadequately prepared for available jobs even though they were fruitfully employed just a few months or years ago.” Does this seem remotely plausible? The supply side plays a role in economic activity, but supply sider logic often puts the supply side through some awfully strange contortions (like cyclical downturns caused by technological regressions) that seem comical until one considers their policy implications.

What are the policy implications of structural unemployment? That demand-side stimuli—expansionary fiscal and monetary policy—can’t help the economy and so shouldn’t be undertaken.

On the other hand, what would need to be true if unemployment were largely “cyclical”? Well, cyclical unemployment arises from an overall deficiency of demand, which is marked by low capacity utilization (check), low inflation (check), low interest rates (check), unemployment spread across many industries (check), and a high ratio of job seekers to job openings across the economy (check). While the presence of these factors doesn’t logically guarantee that unemployment is mainly cyclical, it doesn’t contradict it. Which is more than can be said about the facts in relation to the structural unemployment story.

This isn’t to say there’s no structural unemployment. There’s always some imperfection in the fit between the labor force and productive capacity. And as high unemployment persists and more workers join the ranks of the long-term unemployed, structural unemployment will rise. But that’s a problem we can start to worry about if unemployment ever again reaches near the “natural rate of unemployment” (currently 5.2 percent, according to the CBO). The problem now is that millions of workers across most industries have been thrown out of work because people aren’t buying the products they normally make.

What could be driving the idea that we’re now experiencing mainly structural unemployment? Here, I think, we need to enter the psychological realm. But I’m unclear what the psychological mechanisms could be. Is it that some economists and commentators just feel they need to take up a contrary position to the one “liberal” economists and policymakers espouse and that motivated the obvious actions of stimulating demand when demand was depressed? Is it that policymakers and conservative ideologues just don’t want anything done about unemployment, and are looking for an excuse not to act? Or a way of bringing Obama down?

I sense that a large chunk of it is simply narrative opportunism. The unemployment rate is still high, an opening for the David Brookss of the world to explain that demand-side stimuli can’t really address our problems, which are deeper, “structural” in nature, as profound minds can grasp.

There’s a coherent supply-side story to tell about downturns (just as there’s a coherent demand-side story). But in this instance it doesn’t fit the facts about inflation, capacity utilization, interest rates, or recent historical events like a bursting housing bubble and a plummeting stock market. And of course it’s not remotely plausible that workers who were gainfully employed in many industries just a few years ago are suddenly incapable of running the nation’s productive capacity.